Around the noise. Chatbot: an empty hype or a new era?
We continue a series of interviews with speakers at the DUMP conference .
Hyp around chat bots does not subside. Why is everyone talking about them? Is it difficult to program them? How to make money on them? Is there any chance chatbots will replace mobile apps? What about people? Our questions were answered by the founder of the ConfBot project Alexey Mogilnikov and the technical director of JetStyle , who accounted for a dozen developed chatbots , Alexey Spiridonov.
- Why did you start chat bots? Why is this such a popular topic now? Everyone is talking about chat bots, coming up with chat bots, trying to invest in chat bots. Why did everyone decide what the future was for?
Alexey Mogilnikov: Lena (my wife and cofounder) organizes IT conferences: RailsClub, PyCon Russia, FPConf and DUMP. And the organizers have such eternal themes: how to make the schedule convenient for participants; how to collect their opinion about the conference; How to organize effective networking?
Somehow we discussed how to solve all this. Nobody really puts mobile applications, it’s always a pity to spend money on booklets, and it’s inconvenient to use them, the site is difficult to use from your phone.
The idea with the bot came precisely because it met all the criteria for solving the problem: easy to connect, convenient to use from the phone, fun interactive format. That is, not “let’s come up with and make some bot”, but “oh, it’s the bot that will perfectly solve our problem.” Well, and then the sacramental "tyzh programmer", and I sat down to write. So ConfBot appeared .
Chatbots have become a popular topic for a very simple reason: the sharp rise of instant messengers over the past 4 years. Now there is more audience in messengers than on social networks. And if 2.5 years ago only the lazy did not laugh at Zuckerberg, who bought WhatsApp for $ 22 billion, now these people are not funny: MAU WhatsApp'a - 1.2 billion people, and the meter continues to tick.
And since there are people in messengers, and these people have needs, you need to somehow interact with them, provide some services, and sell something. Chatbots are a natural way to do this, right where people are comfortable - in their favorite messenger.
Hyp associated with chatbots will pass, but chatbots will remain. It seems to me that we have already seen enough hyps to roughly guess what will happen next. A direct analogy is hype to mobile applications 5-7 years ago. Everyone did mobile applications. The mobile application was the answer to any question. Some, of course, have not yet been released, but the hype is over and the mobile applications remain. And they will be with us for a very long time. It will be the same with chatbots. This will be another way to interact with users. With its pros and cons.
Alexey Spiridonov: Today everyone has a smartphone and at least one installed messenger, and chat has become more familiar than SMS and mail.
Previously, people sat on sites and searched for information through search engines. Then most of these functions moved to social networks. We got acquainted there, shared information, made an appointment. Now we spend more time in messengers.
It is logical to assume that they will become a new channel for providing services and solving problems that websites or social networks are doing worse.
At JetStyle, we always experiment a lot with what technology development offers us - whether it's chat bots or augmented and virtual reality. Now is the time to accumulate expertise and try new technologies, integrate them into customers' business processes. That is what we are doing.
- Why talk about chat bots, but use little?
Alexey Mogilnikov:They say because hype :) They use it a little, because we are only at the very beginning of the journey. Plus, "conversational chatbots" (chatbots that communicate in natural language) do not work well. Interface chat bots (bots that interact with the user through the user interface: buttons, pictures, webview and other elements) are much more convenient, but only Telegram and the Facebook Messenger have a more or less serious platform for interface chat bots. No matter how we love Telegram, its MAU is 170 million people, this is quite a bit (compare with WhatsApp). Everyone has a Facebook Messenger, but most users do not like it and use it a little. The game changer will be a platform for front-end chat bots from WhatsApp, I believe that sooner or later it will appear.
And also, perhaps, the reason is that chatbots are still quite a “geeky” topic, “ordinary” users are not used to them yet. But it is a matter of time.
Alexey Spiridonov: If it seems to you that chat bots are not used much, then you are just not looking there. For example, in China, out of 900 million people with WeChat installed, about 800 million use bots. For many Chinese, this is almost the entire Internet. Almost everything is done through WeChat: goods are sold, gifts are given, they are recorded in a clinic, a plumber is called, etc.
With regard to Russia, it is worth noting a few points, including the availability of API popular messengers for developers. It seems that as soon as WhatsApp opens the API, the number of bots will increase, because WhatsApp is most popular with us.
Yes, in 2016 there was a hype associated with chat bots. Now it is falling, technology is moving along the Gartner curve further to the abyss of disappointment. And then, if the chatbots justify themselves, we will come to the widespread use and applicability. But those who are now engaged in the development of chat bots and are looking for their place in their business processes will be the first to benefit.
In January 2017, Alexey and Elena presented the bot at the Eurasia Mobile Challenge and won it bypassing 1000 other projects
- Now many have the illusion that absolutely everyone needs chat bots. Why do they fit well, but why not?
Alexey Mogilnikov: They really need everyone, just not everyone else knows about it :) And the chat bots themselves have not reached the level of maturity so that everyone can use them.
Now it makes sense to use them for some time-limited small tasks: the first line of support, some background information, an auxiliary channel for ordering food, and more.
Alexey Spiridonov: You need to understand: chat bots are needed not by absolutely everyone, but by those whose user cases contain dialogs. If your tasks are solved through communication with customers, then messengers are well suited to you as a communication channel, and chat bots will complement them. Therefore, the answer to the question: “What can I instruct the chatbot today?” Is very simple: “Everything that can be arranged as a dialogue.”
If we talk about our cases, then the chat bot, for example, is well suited for ordering food, it helps to place an order and arrange delivery, sends notifications about the status of the order. With the help of the chat bot, we began to save quite substantial amounts that our client had previously spent on sending SMS notifications.
- In what area can chatbots and instant messengers really make money in the near future?
Alexey Mogilnikov: Here you need to divide: messengers and chat bots.
Messengers in the near future will not earn serious money, and they don’t need to. The mechanics of a business with products with billions of audiences is a little different. Most likely, something will slightly change when payments finally come to the messengers and they will sit on the flow of money between businesses and their customers. There is little to wait, the Facebook Messenger is already testing payments in the USA. It has been rumored for a year that Telegram plans to tighten payments. VimpelCom’s recently announced new Veon messenger will support payments. In general, the caravan is on, but not fast.
As for chat bots, they are already making money. Not all and a little, but the process goes on as the general development of chat bots and people get used to them, there will obviously be money.
Most likely, by analogy with the gold rush (where brothel and saloon owners primarily earned money, rather than gold diggers), chatbot designers such as Chatfuel, ManyChat and Botmother will earn money in the very near future.
Chatbots that automate routine tasks for business will “earn”, saving business money. The first technical support line is a great example.
E-commerce through chatbots is already feeling pretty good and will feel better and better. For example, a chatbot for Facebook Messenger from Shopify is clearly taking off.
Alexey Spiridonov:Chatbots are making money right now, just maybe not to the extent and extent we would like. We have a client for whom the chat bot earns millions of rubles. Nevertheless, yes, this is just an additional channel. You need to understand this.
As for messengers, the question is more complicated. And I don’t have a clear answer now.
It can be assumed that instant messengers will take a commission for payments within themselves. Perhaps they will earn money from developers - for example, to take a fee for access to advanced features, hosting, user limits, development licenses or something like that. But this is now nothing more than speculation. You need to wait for the actions of the messengers themselves
- Do you believe that chat bots will kill mobile applications?
Alexey Mogilnikov:They will not kill them, but they will squeeze. Mobile applications are a rather difficult story, they need to be downloaded and installed, they are an eyesore, take up space on the phone. Chat bots in this sense are much easier: I quickly connected via the link, got what you need and forgot about it. If spamming, then simply deleted. Less “friction” when connected - more conversion.
In general, mobile apps are in some kind of saturation crisis. There are statistics in the United States that the average number of new mobile applications that users install on their phones is gradually going to zero. Those. people just don’t need new applications, all their basic needs are satisfied, and I don’t really want to put a new application for rare or one-time tasks. This is the place where chatbots will crowd out mobile apps.
Alexey Spiridonov: I think it's too early to say that chat bots will kill mobile applications. Today, application functionality far exceeds the capabilities of bots. But they have a different range of tasks to be solved.
But the fact that mobile applications will become smaller over time is yes. I think that users are full of applications: they need to be installed, they require a lot of Internet, slow down on not the coolest phones, sometimes personal data leaks through them, you need to switch between them ... the list can be continued for a long time.
- Will chat bots be able to replace live workers in the future?
Alexey Mogilnikov:Yes of course. This is not even a matter of chatbots, but the fact that software more and more can save humanity from routine tasks. This means that people who solve these routine tasks will have to look for themselves in new areas. More intelligent or those we have not yet reached - programmers (mu ha ha).
Alexey Spiridonov: I think that over time, chatbots will be able to replace live workers in some areas, for example, in technical support. Already there are successful cases.
But here we go into the territory of machine learning and neural networks, for which huge arrays of sorted data are needed. Few can afford this now - either because they have never collected such data, or because it is a long and expensive process. Therefore, chatbots will still work in close conjunction with the person.
One of the successful JetStyle cases under the technical direction of Alexey Spiridonov is a bot for the Papa John's Pizza pizzeria chain
. Do you need to be a programmer to create a chat bot? What are the difficulties in the development?
Alexey Mogilnikov: No, it is not necessary to be a programmer. There are chatbot constructors that allow you to create a chatbot through the user interface by clicking with the mouse. More or less complex bots, of course, need to be programmed.
There are no serious difficulties in the development of chat bots. These are regular web applications without an html frontend. There are a number of nuances, but they are comprehended with experience.
Alexey Spiridonov: It all depends on why you need a bot. If you have enough simple - use the constructor, they are now.
We are moving towards the integration of chat bots with the customer’s internal systems, so programming is indispensable here. Moreover, not only a programmer is needed, the participation of an interface designer and copywriter is extremely important. Our experience shows that it is too early to start programming without preliminary scripting - then redo it.
- Your forecast: how will chat bots develop in the next two years?
Alexey Mogilnikov:First of all, I believe that the fashion for conversational chatbots will pass. These are chat bots that try to communicate with the user in a natural language. And it turns out pretty bad for them due to the complexity of recognizing user input. Plus, it's all pretty uncomfortable. It is commonplace for a user to press a button more conveniently and quickly than to type text on the on-screen keyboard of a smartphone. Perhaps the future lies with hybrid chatbots that recognize user input wherever convenient, but they solve all other tasks with an interface.
I believe that in the next two years, more or less large messengers will provide their platforms for chat bots, payments will appear in them, and chat bots will become another, rather good, channel for interaction between business and customers. They will also become a separate, albeit rather specific, form of software that helps users solve their daily tasks.
Add this post to bookmarks and check in two years :)
Alexey Spiridonov: If we are talking about full-fledged interlocutors on any topic, then chat bots in Russian are at the very beginning of the way. And I think that breakthrough solutions will not appear in the next two years.
If we are talking about “interface”, simpler bots, then changes will certainly occur. Almost every month, leading messengers launch new features (while we have been analyzing the capabilities of different messengers even the “What's New in Bots” summary table has been compiled and updated), the appearance patterns of chat bots are formed, there will be more designers for chat bots, more hosting sites designed for them, I think that there will be more libraries or frameworks in popular languages for working with chat bots.
But, as I said, now you need to experiment, accumulate expertise and look for a place for chat bots in business processes.
On April 14, at the DUMP conference, Alexey Mogilnikov will show on a simple example how to write an interesting chat bot. The example will be in Ruby, but chatbots are a story in which the language is even less important than usual. So, even if you are not a hack, it will still be clear and useful. Have questions and come listen.
Thanks to our sponsors: the general sponsor is E-Soft , the conference partners are SKB Kontur , Naumen , Sberbank Technologies .
Hyp around chat bots does not subside. Why is everyone talking about them? Is it difficult to program them? How to make money on them? Is there any chance chatbots will replace mobile apps? What about people? Our questions were answered by the founder of the ConfBot project Alexey Mogilnikov and the technical director of JetStyle , who accounted for a dozen developed chatbots , Alexey Spiridonov.
- Why did you start chat bots? Why is this such a popular topic now? Everyone is talking about chat bots, coming up with chat bots, trying to invest in chat bots. Why did everyone decide what the future was for?
Alexey Mogilnikov: Lena (my wife and cofounder) organizes IT conferences: RailsClub, PyCon Russia, FPConf and DUMP. And the organizers have such eternal themes: how to make the schedule convenient for participants; how to collect their opinion about the conference; How to organize effective networking?
Somehow we discussed how to solve all this. Nobody really puts mobile applications, it’s always a pity to spend money on booklets, and it’s inconvenient to use them, the site is difficult to use from your phone.
The idea with the bot came precisely because it met all the criteria for solving the problem: easy to connect, convenient to use from the phone, fun interactive format. That is, not “let’s come up with and make some bot”, but “oh, it’s the bot that will perfectly solve our problem.” Well, and then the sacramental "tyzh programmer", and I sat down to write. So ConfBot appeared .
Chatbots have become a popular topic for a very simple reason: the sharp rise of instant messengers over the past 4 years. Now there is more audience in messengers than on social networks. And if 2.5 years ago only the lazy did not laugh at Zuckerberg, who bought WhatsApp for $ 22 billion, now these people are not funny: MAU WhatsApp'a - 1.2 billion people, and the meter continues to tick.
And since there are people in messengers, and these people have needs, you need to somehow interact with them, provide some services, and sell something. Chatbots are a natural way to do this, right where people are comfortable - in their favorite messenger.
Hyp associated with chatbots will pass, but chatbots will remain. It seems to me that we have already seen enough hyps to roughly guess what will happen next. A direct analogy is hype to mobile applications 5-7 years ago. Everyone did mobile applications. The mobile application was the answer to any question. Some, of course, have not yet been released, but the hype is over and the mobile applications remain. And they will be with us for a very long time. It will be the same with chatbots. This will be another way to interact with users. With its pros and cons.
Alexey Spiridonov: Today everyone has a smartphone and at least one installed messenger, and chat has become more familiar than SMS and mail.
Previously, people sat on sites and searched for information through search engines. Then most of these functions moved to social networks. We got acquainted there, shared information, made an appointment. Now we spend more time in messengers.
It is logical to assume that they will become a new channel for providing services and solving problems that websites or social networks are doing worse.
At JetStyle, we always experiment a lot with what technology development offers us - whether it's chat bots or augmented and virtual reality. Now is the time to accumulate expertise and try new technologies, integrate them into customers' business processes. That is what we are doing.
- Why talk about chat bots, but use little?
Alexey Mogilnikov:They say because hype :) They use it a little, because we are only at the very beginning of the journey. Plus, "conversational chatbots" (chatbots that communicate in natural language) do not work well. Interface chat bots (bots that interact with the user through the user interface: buttons, pictures, webview and other elements) are much more convenient, but only Telegram and the Facebook Messenger have a more or less serious platform for interface chat bots. No matter how we love Telegram, its MAU is 170 million people, this is quite a bit (compare with WhatsApp). Everyone has a Facebook Messenger, but most users do not like it and use it a little. The game changer will be a platform for front-end chat bots from WhatsApp, I believe that sooner or later it will appear.
And also, perhaps, the reason is that chatbots are still quite a “geeky” topic, “ordinary” users are not used to them yet. But it is a matter of time.
Alexey Spiridonov: If it seems to you that chat bots are not used much, then you are just not looking there. For example, in China, out of 900 million people with WeChat installed, about 800 million use bots. For many Chinese, this is almost the entire Internet. Almost everything is done through WeChat: goods are sold, gifts are given, they are recorded in a clinic, a plumber is called, etc.
With regard to Russia, it is worth noting a few points, including the availability of API popular messengers for developers. It seems that as soon as WhatsApp opens the API, the number of bots will increase, because WhatsApp is most popular with us.
Yes, in 2016 there was a hype associated with chat bots. Now it is falling, technology is moving along the Gartner curve further to the abyss of disappointment. And then, if the chatbots justify themselves, we will come to the widespread use and applicability. But those who are now engaged in the development of chat bots and are looking for their place in their business processes will be the first to benefit.
In January 2017, Alexey and Elena presented the bot at the Eurasia Mobile Challenge and won it bypassing 1000 other projects
- Now many have the illusion that absolutely everyone needs chat bots. Why do they fit well, but why not?
Alexey Mogilnikov: They really need everyone, just not everyone else knows about it :) And the chat bots themselves have not reached the level of maturity so that everyone can use them.
Now it makes sense to use them for some time-limited small tasks: the first line of support, some background information, an auxiliary channel for ordering food, and more.
Alexey Spiridonov: You need to understand: chat bots are needed not by absolutely everyone, but by those whose user cases contain dialogs. If your tasks are solved through communication with customers, then messengers are well suited to you as a communication channel, and chat bots will complement them. Therefore, the answer to the question: “What can I instruct the chatbot today?” Is very simple: “Everything that can be arranged as a dialogue.”
If we talk about our cases, then the chat bot, for example, is well suited for ordering food, it helps to place an order and arrange delivery, sends notifications about the status of the order. With the help of the chat bot, we began to save quite substantial amounts that our client had previously spent on sending SMS notifications.
- In what area can chatbots and instant messengers really make money in the near future?
Alexey Mogilnikov: Here you need to divide: messengers and chat bots.
Messengers in the near future will not earn serious money, and they don’t need to. The mechanics of a business with products with billions of audiences is a little different. Most likely, something will slightly change when payments finally come to the messengers and they will sit on the flow of money between businesses and their customers. There is little to wait, the Facebook Messenger is already testing payments in the USA. It has been rumored for a year that Telegram plans to tighten payments. VimpelCom’s recently announced new Veon messenger will support payments. In general, the caravan is on, but not fast.
As for chat bots, they are already making money. Not all and a little, but the process goes on as the general development of chat bots and people get used to them, there will obviously be money.
Most likely, by analogy with the gold rush (where brothel and saloon owners primarily earned money, rather than gold diggers), chatbot designers such as Chatfuel, ManyChat and Botmother will earn money in the very near future.
Chatbots that automate routine tasks for business will “earn”, saving business money. The first technical support line is a great example.
E-commerce through chatbots is already feeling pretty good and will feel better and better. For example, a chatbot for Facebook Messenger from Shopify is clearly taking off.
Alexey Spiridonov:Chatbots are making money right now, just maybe not to the extent and extent we would like. We have a client for whom the chat bot earns millions of rubles. Nevertheless, yes, this is just an additional channel. You need to understand this.
As for messengers, the question is more complicated. And I don’t have a clear answer now.
It can be assumed that instant messengers will take a commission for payments within themselves. Perhaps they will earn money from developers - for example, to take a fee for access to advanced features, hosting, user limits, development licenses or something like that. But this is now nothing more than speculation. You need to wait for the actions of the messengers themselves
- Do you believe that chat bots will kill mobile applications?
Alexey Mogilnikov:They will not kill them, but they will squeeze. Mobile applications are a rather difficult story, they need to be downloaded and installed, they are an eyesore, take up space on the phone. Chat bots in this sense are much easier: I quickly connected via the link, got what you need and forgot about it. If spamming, then simply deleted. Less “friction” when connected - more conversion.
In general, mobile apps are in some kind of saturation crisis. There are statistics in the United States that the average number of new mobile applications that users install on their phones is gradually going to zero. Those. people just don’t need new applications, all their basic needs are satisfied, and I don’t really want to put a new application for rare or one-time tasks. This is the place where chatbots will crowd out mobile apps.
Alexey Spiridonov: I think it's too early to say that chat bots will kill mobile applications. Today, application functionality far exceeds the capabilities of bots. But they have a different range of tasks to be solved.
But the fact that mobile applications will become smaller over time is yes. I think that users are full of applications: they need to be installed, they require a lot of Internet, slow down on not the coolest phones, sometimes personal data leaks through them, you need to switch between them ... the list can be continued for a long time.
- Will chat bots be able to replace live workers in the future?
Alexey Mogilnikov:Yes of course. This is not even a matter of chatbots, but the fact that software more and more can save humanity from routine tasks. This means that people who solve these routine tasks will have to look for themselves in new areas. More intelligent or those we have not yet reached - programmers (mu ha ha).
Alexey Spiridonov: I think that over time, chatbots will be able to replace live workers in some areas, for example, in technical support. Already there are successful cases.
But here we go into the territory of machine learning and neural networks, for which huge arrays of sorted data are needed. Few can afford this now - either because they have never collected such data, or because it is a long and expensive process. Therefore, chatbots will still work in close conjunction with the person.
One of the successful JetStyle cases under the technical direction of Alexey Spiridonov is a bot for the Papa John's Pizza pizzeria chain
. Do you need to be a programmer to create a chat bot? What are the difficulties in the development?
Alexey Mogilnikov: No, it is not necessary to be a programmer. There are chatbot constructors that allow you to create a chatbot through the user interface by clicking with the mouse. More or less complex bots, of course, need to be programmed.
There are no serious difficulties in the development of chat bots. These are regular web applications without an html frontend. There are a number of nuances, but they are comprehended with experience.
Alexey Spiridonov: It all depends on why you need a bot. If you have enough simple - use the constructor, they are now.
We are moving towards the integration of chat bots with the customer’s internal systems, so programming is indispensable here. Moreover, not only a programmer is needed, the participation of an interface designer and copywriter is extremely important. Our experience shows that it is too early to start programming without preliminary scripting - then redo it.
- Your forecast: how will chat bots develop in the next two years?
Alexey Mogilnikov:First of all, I believe that the fashion for conversational chatbots will pass. These are chat bots that try to communicate with the user in a natural language. And it turns out pretty bad for them due to the complexity of recognizing user input. Plus, it's all pretty uncomfortable. It is commonplace for a user to press a button more conveniently and quickly than to type text on the on-screen keyboard of a smartphone. Perhaps the future lies with hybrid chatbots that recognize user input wherever convenient, but they solve all other tasks with an interface.
I believe that in the next two years, more or less large messengers will provide their platforms for chat bots, payments will appear in them, and chat bots will become another, rather good, channel for interaction between business and customers. They will also become a separate, albeit rather specific, form of software that helps users solve their daily tasks.
Add this post to bookmarks and check in two years :)
Alexey Spiridonov: If we are talking about full-fledged interlocutors on any topic, then chat bots in Russian are at the very beginning of the way. And I think that breakthrough solutions will not appear in the next two years.
If we are talking about “interface”, simpler bots, then changes will certainly occur. Almost every month, leading messengers launch new features (while we have been analyzing the capabilities of different messengers even the “What's New in Bots” summary table has been compiled and updated), the appearance patterns of chat bots are formed, there will be more designers for chat bots, more hosting sites designed for them, I think that there will be more libraries or frameworks in popular languages for working with chat bots.
But, as I said, now you need to experiment, accumulate expertise and look for a place for chat bots in business processes.
On April 14, at the DUMP conference, Alexey Mogilnikov will show on a simple example how to write an interesting chat bot. The example will be in Ruby, but chatbots are a story in which the language is even less important than usual. So, even if you are not a hack, it will still be clear and useful. Have questions and come listen.
Thanks to our sponsors: the general sponsor is E-Soft , the conference partners are SKB Kontur , Naumen , Sberbank Technologies .